Effects of L-Dopa and Oxazepam on Resting-State Functional Magnetic Resonance Imaging Connectivity: A Randomized,Cross-Sectional Placebo Study

Abstract

Pharmacological functional brain imaging has traditionally focused on neuropharmacological modulations of event-related responses. The current study is a randomized, cross-sectional resting-state functional magnetic resonance imaging study where a single dose of commonly prescribed amounts of either benzodiazepine (oxazepam), L-dopa, or placebo was given to 81 healthy subjects. It was hypothesized that the connectivity in resting-state networks would be altered, and that the strength of connectivity in areas rich in target receptors would be particularly affected. Additionally, based on known anxiolytic mechanisms of benzodiazepines, modulated amygdala (Am) connectivity was predicted. To test this, seed region-based correlational analysis was performed using seven seeds placed in well-characterized resting-state networks, in regions with above-average densities of GABA-A or dopamine receptors and in Am. To alleviate the anatomical bias introduced by the a priori selected seed regions, whole-brain exploratory analysis of regional homogeneity and fractional amplitude of low-frequency fluctuations (fALFF) was also carried out. Oxazepam increased functional connectivity between midline regions of the default-mode network (DMN) and the prefrontal, parietal, and cerebellar areas, but decreased connectivity between, for example, the Am and temporal cortex. L-dopa mainly decreased connectivity between the Am and bilateral inferior frontal gyri and between midline regions of the DMN. The fALFF analysis revealed that L-dopa decreased low-frequency fluctuations in the cerebellum. It was concluded that the overall effects of single administrations of oxazepam and L-dopa on resting-state connectivity were small both in strength and in spatial extent, and were on par with placebo effects as revealed by comparing the two placebo groups.

Introduction

P harmacological functional magnetic resonance imaging (ph-fMRI) is a noninvasive method traditionally used for studying the effects of drugs on task-evoked brain activity (for a review, see Honey and Bullmore, 2004). Recently, functional connectivity analysis of resting-state functional magnetic resonance imaging (rs-fMRI) data has emerged as a promising tool to map neuronal circuitry (Fox and Raichle, 2007) as well as for characterizing deviant resting-state networks among clinical populations (reviewed in Rosazza and Minati, 2011). Additionally, rs-fMRI studies have successfully revealed modulations of brain connectivity related to psychoactive substances (e.g., Mccabe et al., 2011). The idea that there might exist an interaction between drug action and the brain's baseline activity is attractive, given that the brain's baseline activity has been estimated to consume the vast majority of the brain's total energy need (Raichle and Mintun, 2006). Moreover, since rs-fMRI is conducted without the need of giving explicit instructions to the participants, variability of data due to subtle differences in task compliance and performance is substantially reduced.

In this study, modulatory effects on rs-fMRI activity from two psychoactive drugs that are commonly used in the clinical setting were focused: benzodiazepine (oxazepam) and dopamine (L-dopa). By investigating two different drugs using the identical methodology and statistical analysis, better estimation of the specificity of the effects of each drug was aimed. Resting-state networks that were a priori hypothesized would be affected mainly by, for example, oxazepam could thus serve as control networks when examining the effects of L-dopa, and vice versa (in accordance with the guidelines advocated by e.g., Fox and Greicius, 2010). Benzodiazepines constitute a class of psychoactive drugs commonly used as anxiolytics and for presurgical sedation. Benzodiazepines inhibit brain activity through GABA-A receptors (Tan et al., 2011) distributed throughout the brain with the highest densities found in the occipital cortex (Abadie et al., 1992). The effects of benzodiazepine (midazolam)-induced sedation on rs-fMRI connectivity have, in a previous study, mainly been associated with increased connectivity in motor and visual areas and decreased connectivity in the default-mode network (DMN) (Greicius et al., 2008). However, to the best of our knowledge, no previous study has investigated the effects on rs-fMRI connectivity from small, nonsedative doses of benzodiazepines typically administered to patients suffering from anxiety disorders.

In an independent subject cohort, the effects of L-dopa on resting-state functional connectivity were studied. L-dopa is a precursor to dopamine, which in contrast to dopamine crosses the blood–brain barrier and therefore is used as a dopaminergic drug, mainly against parkinsonism. Dopamine receptors are widespread throughout the brain, with highest densities of the most common types (DA1 and DA2) found in the striatum and the brainstem (Volkow et al., 1996). The majority of previous studies on dopaminergic modulation of fMRI connectivity have focused on event-related changes or changes in intrinsic activity in clinical populations. These studies have reported that L-dopa can both increase and decrease connectivity (Honey et al., 2003; Kwak, 2010). An important study by Kelly and colleagues, investigating the effects of L-dopa on resting-state activity in healthy subjects, found increased connectivity between putamen (Put) and the cerebellum, between the nucleus accumbens (NA) and the ventrolateral prefrontal cortex, as well as a decreased connectivity between the caudate nucleus and the DMN (Kelly et al., 2009). However, that study was limited in scope to only detect changes in connectivity between prespecified regions in the striatum and the rest of the brain.

In the present study, commonly prescribed doses of oxazepam and L-dopa that earlier have been shown to modulate behavioral outcomes (e.g., Gospic et al., 2011; Kischka et al., 1996) were administered. Three hypotheses were posed: Oxazepam would primarily alter resting-state networks, encompassing brain regions known to have a high GABA-A receptor density such as the primary visual cortex (PVC). Second, oxazepam would modulate connectivity in regions implicated in anxiety disorders, such as amygdala (Am). Finally, L-dopa would primarily alter resting-state networks residing in areas with a high dopamine receptor density such as striatum, but less so for networks with a relatively lower DA receptor density.

Methods

Subjects

Eighty-one right-handed volunteers divided into four groups were included in this study. Subjects in the first phase of data collection were randomly assigned to either the oxazepam group (n=20, 11 women, 24.6±4.4 years) or the oxazepam placebo group (n=22, 14 women, 25.5±4.6 years). In a second phase of data collection, subjects were randomly assigned to either the L-dopa group (n=19, 9 women, 22.3±3.5 years) or the L-dopa placebo group (n=20, 10 women, 22.8±4.5). There were no significant group differences with regard to gender or age between any of the compared groups, that is, between the oxazepam group and the oxazepam placebo group (x² [1, n=42]=0.32, p=0.67; t[40]=0.68, p=0.50), between the L-dopa group and the L-dopa placebo group (x² [1, n=39]=0.02, p=0.89; t[37]=0.38, p=0.71), between the oxazepam placebo and the L-dopa placebo (x² [1, n=42]=0.80, p=0.37; t[40]=1.95, p=0.059), or between the oxazepam group and the L-dopa group (x² [1, n=40]=0.10, p=0.75; t[37]=1.81, p=0.078). All subjects were healthy, took no medications, and had no prior or present history of psychiatric illness or neurological disease. Participants were recruited at the campus of the Karolinska Institute and gave informed consent for participation. The study was approved by the local governmental ethics committee in Stockholm, Sweden.

rs-fMRI paradigm

The resting-state data analyzed in the current study were collected at two separate occasions in addition to event-related fMRI sessions reported elsewhere (Gospic et al., 2011; Gospic et al., in preparation). The first study investigated the effects of oxazepam in a decision-making paradigm, and the second study that was conducted about 12 months later studied the effects of L-dopa for the same decision-making paradigm (event-related fMRI designs). The experimental procedures used in the two studies were identical except for the drug given, the information given to the participants on their respective effects (see below), and that the L-dopa study was double blinded, whereas the oxazepam study was blinded only for the participants. Upon arrival, participants were randomly assigned to either the placebo group or the drug group. Subjects in the oxazepam group received a pill containing 20 mg of oxazepam (Sobril). Subjects in the L-dopa group were administered a pill containing 100 mg L-dopa and 25 mg of benserazide (a decarboxylase blocker preventing peripheral side effects; Madopark). In the oxazepam study, subjects were told that they would receive either placebo or “… a benzodiazepine agonist that has a calming effect. Higher doses can make you sleepy and reduce breathing frequency…,” and subjects in the L-dopa study were told that they would receive either placebo or a “… stimulating drug.…[that] could cause side-effects such as nausea, vomiting, confusion and hallucinations.” The subjects had been asked not to eat for 2 h before the experiment or drink alcohol for 24 h before the experiment. Resting-state data were acquired after the event-related fMRI scans, at ∼90 min after drug administration.

MR imaging acquisition

Functional imaging data were acquired using a 1.5T GE scanner. A T2*-weighted echo planar images sequence (TR/TE=2500/40 ms) sensitized to the blood oxygenation level-dependent (BOLD) contrast was used to obtain 200 image volumes during the resting-state. Each image volume contained 32 slices (4.5-mm thickness, slice gap=0.5 mm, image matrix=64×64 voxels) with a field of view of 220×220 mm. The rs-MR image acquisition lasted for 8 min during which the subjects were instructed to lie still and rest, with their eyes directed at a white fixation cross.

Image preprocessing

MatLab (MathWorks, Inc., Natick, MA) and the toolbox SPM8 (www.fil.ion.ucl.ac.uk/spm/software/) were used for image preprocessing. First, each image volume was realigned to the subject's mean image volume. Second, the realigned volumes were normalized to the Montreal Neurological Institute (MNI) template provided within SPM8 and resampled to cubic voxels (3×3×3 mm³). Finally, image volumes were spatially filtered with a Gaussian kernel with a full-width half-maximum (FWHM) of 6 mm.

Connectivity analysis

rs-fMRI connectivity analysis is currently conducted using a wide range of analytical approaches. Here, the commonly used method of seed-based region-of-interest (ROI) correlation analysis, a hypothesis-driven technique for detecting functional correlation between a priori selected ROIs and the remaining brain, was employed. The bias introduced by the manual selection of ROIs was addressed by including two separate exploratory analyses: fractional amplitude of low-frequency fluctuations (fALFF) (Zou et al., 2008) and coherence regional homogeneity (Cohe-ReHo) (Liu et al., 2010). fALFF is a measure of the ratio of the ALFFs of interest (0.01–0.08 Hz) (see Biswal et al., 1995) to the amplitude of oscillations in the full frequency range. Cohe-ReHo is a measure of the degree to which the BOLD signals in neighboring voxels covary, thus measuring local connectivity. A high Cohe-ReHo value in one voxel implies a high degree of low-frequency coherence between its time series and the time series of its neighbors. A measure of ReHo has, in previous studies, been used to characterize, for example, the effects of long-term treatments with L-dopa (Wu et al., 2009). Both the fALFF and the Cohe-ReHo analysis render parameter estimates for every voxel in the brain.

It is well known that BOLD signal intensity time course contains signal variance unrelated to changes in neuronal activity (Van Dijk et al., 2010). Therefore, in our ROI-based analysis, nine nuisance regressors of no-interest were included in the general linear model. Six regressors modeled residual subject movement (three translational and three rotational parameters), and the remaining three modeled signal variance related to the white matter, cerebrospinal fluid, and the overall global brain signal. The BOLD signal intensity time series from the white matter and from the cerebrospinal fluid were extracted using the Marsbar (http://marsbar.sourceforge.net/) MatLab toolbox by manually defining spherical ROIs (radius=6 mm) at [−32, −45, 29] (white matter) and [23, −37, 16] (cerebrospinal fluid) in the MNI space (analogous to Van Dijk et al., 2010).

Given the growing concern regarding the influence of residual subject motion on rs-fMRI connectivity (i.e., Satterthwaite et al., 2012; Van Dijk et al., 2010), the frame-wise displacement was calculated according to the scheme described in Power and associates (2012) at the subject level, and compared across subject groups. However, no significant group difference of head movement was found (F[3,80]=0.19, p=0.91). No subject movements exceeded the voxel size (3 mm).

For the seed-based ROI correlation analysis, BOLD signal time courses were extracted in seven ROIs (spherical, radius=6 mm) positioned in the left hemisphere, in the brain areas that previously have been shown to belong to commonly observed resting-state networks (Damoiseaux et al., 2006; Di Martino et al., 2008; Fox et al., 2005; Fransson, 2005; Greicius et al., 2003; van den Heuvel and Hulshoff, 2010). Specifically, ROIs were located in the (1) posterior cingulate cortex (PCC, [x,y,z]=[−5, −53, 40]); (2) ventromedial prefrontal cortex (vmPFC, [−1, 47, −4]); (3) primary motor cortex (M1, [−45, −23, 50]); (4) PVC [−15, −100, 0]; (5) NA [−10, 10, 10]; (6) Put [−20, 15, 0]; and (7) Am [−25, 1, −18]. These ROIs were chosen on the basis to investigate whether oxazepam and L-dopa alter connectivity in the DMN and in primary motor networks that often are implicated in the dopamine-depleted state of parkinsonism. Additionally, the resting-state network in the PVC was an important target due to its high concentration of GABA-A receptors. The Am was also included due to its involvement in anxiety disorders (Rauch et al., 2003) and response to benzodiazepine administration (Del-Ben et al., 2010; Paulus et al., 2005). Finally, two striatal ROIs in NA and Put were included due to their high dopamine receptor densities, similar to Kelly and associates (2009). The center coordinates of the ROIs were selected based on earlier work (PCC and vmPFC corresponded to coordinates employed in Fox et al., 2005; NA and Put corresponded to coordinates in Kelly et al., 2009), and then manually verified using the automated anatomical labeling (AAL) atlas provided in MRIcron software (www.mccauslandcenter.sc.edu/mricro/mricron/). The remaining ROIs (M1, PVC, and Am) were manually picked using the AAL atlas. Since there was no a priori hypothesis of left-right asymmetries with regard to functional connectivity changes, the ROIs were restricted to the left hemisphere. We strived to keep the number of the seed–voxel correlational analyses to a necessary minimum that allowed us to test our hypothesis without inflating the number of false positives, in a similar vein to previous investigations (see e.g., Fox et al., 2005; Kelly et al., 2009). At a subject level, seven linear regressions were performed, one for each ROI, using the general linear model formalism as implemented in SPM8. Each model included nine covariates of no-interest as described above, together with a band-passed version of the mean signal time series extracted in the ROI (pass-band: 0.012–0.1 Hz). Only positive correlations were analyzed due to the ongoing controversy regarding the extent to which anticorrelations are artificially induced when controlling for the global mean brain signal (Murphy et al., 2009; Weissenbacher et al., 2009). Subsequently, the resulting subject-specific activation maps were entered into second-level analysis for comparisons of strength of correlations across the placebo and drug group, using two-sample t-tests. The effects of age and sex were controlled by including these as nuisance variables in the general linear model.

The exploratory whole-brain analysis of low-frequency BOLD signal fluctuations was performed on the same realigned, normalized, and smoothed images used for the seed-based ROI correlation analysis. The ratio between the ALFFs (0.01–0.08 Hz) and the amplitude for the entire sampled frequency range (0–0.25 Hz) was computed for each brain voxel at the subject level using REST MatLab toolbox software (http://restfmri.net). The voxel-based ratio computation yielded the so-called fractional ALFF maps. Next, the subject-level fALFF maps were normalized by division with the mean fALFF value of the entire brain, so that fALFF values larger than one implied that the amplitude of local signal fluctuations was above the brain average.

Measures of local connectivity were calculated on realigned, normalized, but not spatially smoothed, images and measured in terms of Cohe-ReHo as implemented in the REST MatLab software package. Of note, the Cohe-ReHo measure of local connectivity was used rather than the related Kendall correlation coefficient (KCC)-ReHo measure, since the former has been shown to produce more valid results in a comparative analysis (Liu et al., 2010). Linear signal trends were removed, and signal intensity time courses were filtered to only contain oscillations in the frequency window of interest (0.01–0.08 Hz). The cluster size that defines what counts as neighboring voxels was set to 27 voxels (default value). Each voxel Cohe-ReHo value were normalized by division with the mean Cohe-ReHo value for the entire brain, and the resulting maps were smoothed using a Gaussian FWHM of 6 mm. Activation maps for all three types of between-group analyses (seed-based ROI correlation analysis, fALFF, and Cohe-ReHo values) were thresholded at p<0.001 uncorrected at a voxel level, using a minimum cluster size of 13 voxels corresponding to a cluster-level-corrected p-value of 0.05, according to Monte Carlo simulations carried out using AlphaSim (cluster-connecting radius rmm=5 mm, 1000 simulations).

Results

As a test of the validity of the analytical methods used, as well as to assure that our ROIs were positioned in locations that adequately represent hubs in resting-state networks as previously described, ROI correlation maps for each ROI were first calculated for each group separately (L-dopa, L-dopa placebo, oxazepam, and oxazepam placebo) as shown in Supplementary Fig. S1a–d (Supplementary Data are available online at www.liebertpub.com/brain). These ROI correlation activation maps were thresholded at p<0.05, family-wise error corrected (FWE). In all the four subject groups, the seed–ROI correlation analyses produced a robust and strong network activity in the DMN (precuneus, PCC, bilateral parietal cortex, and medial prefrontal cortex), bilateral primary motor cortices, visual areas, bilateral striatum, or bilateral Am, respectively. Further, fALFF maps computed within the group showed an above-average presence of spontaneous, low-frequency fluctuations in the brain cortex, with particularly high amplitudes found in the brain areas belonging to the DMN (precuneus/PCC and the bilateral parietal cortex; Supplementary Fig. S2). This observation is in agreement with previous results that have shown above-average intensities of low-frequency signal changes in the DMN (Fransson, 2005). Taken together, the within-group results provided support to the validity of the anatomical localization of the selected ROIs, and that the within-group sensitivity to resting-state connectivity was high.

Next, the putative differences in connectivity between groups were assessed by two-sample t-tests. The effects of oxazepam and L-dopa on all estimated parameters (seed-based ROI correlation analysis, fALFF, and Cohe-ReHo) were generally found to be small. Nevertheless, significant changes were detected. Oxazepam increased functional connectivity between ROIs positioned in the DMN and caudate nucleus, parietal cortex, visual cortex, midline prefrontal areas, and cerebellum, compared to the corresponding placebo group (Fig. 1). Specifically, the oxazepam group showed an increased connectivity between the vmPFC seed and the right supramarginal sulcus, the proximity of the left precentral sulcus, the left calcarine sulcus, the right middle cingulate cortex, and the right cerebellum. The PCC seed showed an increased connectivity with a region near the left caudate nucleus and the left inferior parietal cortex. Additionally, seed regions positioned in the Put and the M1 showed an increased connectivity with the left cerebellum and the prefrontal cortex, respectively. The only detected decreases in connectivity for the oxazepam group with respect to placebo were between the seed in PVC and the supplementary motor areas, the NA seed and the left olfactory bulb, and between the Am seed and left middle temporal gyrus.

FIG. 1.

Pharmacologically induced changes in functional connectivity. Elevated functional connectivity in the drug groups relative the placebo groups are depicted with blue lines. Red lines indicate decreases. The volumes of the yellow spheres (seed regions), the red (clusters with decreased connectivity), and the blue (clusters with increased connectivity) are proportional to the number of voxels contained in the region-of-interests (ROIs) and activation clusters. Left: Oxazepam was mainly associated with increased functional connectivity outside well-characterized resting-state networks, particularly between the ventromedial prefrontal cortex (vmPFC) seed and the remaining brain. However, oxazepam also decreased functional connectivity, notably between the amygdala (Am) and the temporal cortex. Right: L-dopa was mainly associated with decreased functional connectivity, for example, between the seed in the primary visual cortex (PVC) and the temporal cortex, between the Am and the bilateral frontal cortex and between nodes in the default-mode network. All connectivity maps were thresholded at p<0.001 uncorrected at a voxel level, using a minimum cluster size of 13 voxels corresponding to a corrected cluster significance level of p<0.05.

In contrast, the L-dopa group showed a general tendency of reduced connectivity relative to placebo. In detail, L-dopa-induced decreases in connectivity were found between the vmPFC seed and the left cuneus, as well as between the PVC seed and the right superior temporal sulcus (Fig. 1). L-dopa also decreased connectivity between the Am seed and bilateral frontal gyri. All between-group differences in connectivity measured using seed-based ROI analysis are described in Table 1.

Table 1.

All Drug-Induced Changes in Connectivity as Revealed by the Seven Seed–Region-of-Interest Correlational Analyses

Contrast	Seed ROI	Anatomical region	T-value	Cluster size	x	y	z
Oxazepam>placebo	PCC	L caudate	4.30	16	−12	26	10
		L inferior parietal lobule	3.88	17	−45	−43	40
	vmPFC	R supramarginal gyrus	4.74	28	54	−25	40
		L precentral sulcus	4.55	39	−21	−19	55
		L calcarine sulcus	4.49	24	−12	−91	−5
		R cerebellum	4.39	36	12	−70	−35
		R cerebellum	4.27	23	30	−55	−44
		R middle cingulum	4.11	25	6	−28	52
	M1	L superior medial frontal gyrus	4.23	25	−3	47	49
	Put	L cerebellum	4.23	37	−30	−85	−29
Placebo>oxazepam	PVC	L supplementary motor area	4.26	15	0	5	64
	NA	L olfactory bulb	3.70	15	−3	17	−8
	Am	L middle temporal sulcus	3.84	13	−54	−13	−17
L-dopa>placebo	NA	L superior frontal sulcus	4.24	15	−18	44	22
Placebo>L-dopa	vmPFC	L cuneus	4.59	30	−12	−73	22
	PVC	R superior temporal gyrus	4.27	20	60	−34	7
	Am	L inferior frontal gyrus	4.71	32	−48	20	16
		R inferior frontal gyrus	4.06	40	51	35	16

p<0.001 uncorrected at a voxel level, using a minimum cluster size of 13 voxels corresponding to a corrected cluster significance level of p<0.05.

NA, nucleus accumbens; ROI, region-of-interest; vmPFC, ventromedial prefrontal cortex; PVC, primary visual cortex; PCC, posterior cingulate cortex; M1, primary motor cortex.

No drug-induced changes in local connectivity were found using the Cohe-ReHo measure, neither for the oxazepam group nor for the L-dopa group. Similarly, local differences in fALFF did not substantially differ between the groups, apart from a localized decrease in the left cerebellum in the L-dopa group compared to the placebo. To test the possibility that local differences in fALFF resulted in long-range spanning changes in functional connectivity, a post hoc seed-based ROI correlation analysis was performed using a seed placed at the peak voxel in the L-dopa-induced decrease of fALFF [−26, −66, −17]. This revealed an L-dopa decreased connectivity between the left cerebellum and the right middle temporal gyrus (see Supplementary Fig. S3).

Since group comparisons revealed only weak and local differences in functional connectivity that could be attributed to drug effects per se, we were interested in whether a difference in information given to the participants alone (i.e., oxazepam placebo vs L-dopa placebo) would have an effect on intrinsic connectivity. Consequently, a post hoc control analysis was performed where the seed-based ROI correlation for the placebo group for L-dopa with the placebo group for oxazepam was compared. This post hoc analysis showed increased connectivity for the L-dopa placebo group compared to the oxazepam placebo between the vmPFC seed and the bilateral hippocampus, as well as the left precuneus (Supplementary Fig. S4). The difference in DMN connectivity was unexpected and motivated, yet another post hoc analysis where the L-dopa group was directly compared with the oxazepam group. This contrast revealed a similar tendency of increased connectivity for the L-dopa group between the vmPFC seed and hippocampus (although only the right hippocampus survived the significance threshold, see Supplementary Fig. S4).

Discussion

Overall, relatively weak modulatory effects of both L-dopa and benzodiazepine on resting-state functional connectivity were observed as compared to within-group connectivity patterns among healthy subjects. Regarding the effect of oxazepam, with the single exception of an oxazepam-induced increase in connectivity between vmPFC seed region and the calcarine sulcus, no oxazepam-induced changes in connectivity were observed in the brain regions that are typically rich in GABA-A receptors, such as the occipital cortex. As predicted, and in line with previous studies showing benzodiazepine-induced reduction in event-related BOLD responses in the Am (Del-Ben et al., 2010; Paulus et al., 2005), oxazepam-induced decrease in connectivity between Am and middle temporal gyrus was found. The Am is implicated in salience attribution of stimuli (Costafreda et al., 2008), and the middle temporal gyrus is associated with, for example, emotional audiovisual information (Park et al., 2010). Speculatively, the calming influence of oxazepam could in part be due to a functional uncoupling between Am and areas coding for percepts. Oxazepam was furthermore associated with increase in connectivity between seed regions located within the DMN (particularly vmPFC, but also PCC) and cortical areas located outside the DMN. Previous literature on the effects of benzodiazepines other than oxazepam on resting-state connectivity is relatively sparse, mainly focusing on sedative doses and is to date not conclusive (for a review, see Nallasamy and Tsao, 2011). Investigations of the sedative aspects of benzodiazepines have shown that slightly sedative doses of midazolam decrease DMN activity in the PCC on one hand (Greicius et al., 2008), and that DMN is preserved under isoflurane anesthesia in monkeys, on the other hand (Vincent et al., 2007). Kiviniemi and associates (2005) found a general increased amplitude of spontaneous low-frequency signal fluctuations in, for example, the visual cortex in midazolam sedation, and Martuzzi and associates (2010) found increased functional connectivity in primary sensory cortical areas also under sevoflurane anesthesia. In contrast, Peltier and associates (2005) found a dose-dependent reduction under sevoflurane anesthesia of functional connectivity maps of motor cortex obtained by using ROIs in motor areas. These divergent findings of modulated activity in DMN and in sensory areas all had in common that they were accompanied with strong changes in the degree of awareness (e.g., anesthesia), which was not the case in the current study. The conclusion that a single administration of a small dosage of oxazepam has limited effects on intrinsic network connectivity is further underlined by the fact that no changes in local connectivity (Cohe-ReHo) nor in low-frequency fluctuations (fALFF) were observed.

In the case of dopamine, the postulated hypothesis that L-dopa administration would in particular alter striatal connectivity was not convincingly supported. Although increased connectivity between the NA and the left frontal superior sulcus for the L-dopa cohort compared to placebo could be observed, no support for a broader interaction between L-dopa and connectivity in brain networks encompassing structures of high dopamine receptor densities was found. It is of interest to compare the results presented here to the findings described in the study by Kelly and associates (2009), for which modulation of resting-state connectivity by administration of L-dopa in healthy subjects was investigated using a set of seed ROIs that partly overlapped with ours (Put and NA) while controlling for the same nuisance variables, for example, the global mean signal (Kelly et al., 2009). Similar to our results, they found L-dopa-induced increases in connectivity between NA and the prefrontal cortex, although slightly different prefrontal regions were reported. Speculatively, the increased frontostriatal connectivity observed here might be responsible for dopamine-induced improvement of cognitive performance (Cools et al., 2009). Along a similar vein, the finding of L-dopa-induced decrease of DMN connectivity (between the vmPFC seed region and the cuneus) is analogous with the findings of Kelly and associates (2009). In contrast, we could not reproduce their finding of an increased cerebellar–striatal connectivity. Instead, an L-dopa-associated decrease in fALFF in the cerebellum and a decreased cerebellar–temporal functional connectivity were found. This discrepancy between our study and the results obtained by Kelly and colleagues (2009) might be attributed to several causes. First, they employed a repeated-measure design that likely yields increased sensitivity. Second, they acquired data at 4 Tesla, which constitute an almost threefold increase in BOLD sensitivity compared to the magnetic field strength used here (1.5 T). Finally, their data were analyzed at a slightly higher spatial resolution, which might have an impact on anatomical ROI specificity.

The observed L-dopa-induced changes in connectivity between Am and frontal gyri resemble earlier findings of dopaminergic influence on Am. Kienast and associates (2008) correlated the dopamine storage capacity in Am with Am response to aversive stimuli, and report a correlation between the strength of the functional connectivity between the Am and anterior cingulate cortex and anxiety trait. In an aversive conditioning paradigm, Diaconescu and associates (2010) found a dopamine-modulated event-related change in connectivity between the Am and the middle frontal gyrus, supporting our finding of a dopamine-sensitive connection between these areas. However, the behavioral significance of such pharmacological manipulation remains to be investigated.

It is noteworthy that the majority of the detected modulations on intrinsic connectivity caused by drug intake occurred between nodes inside the targeted resting-state networks and regions located elsewhere, rather than between the brain areas within the boundaries of the established anatomical definitions of known rs-fMRI networks (shown in Supplementary Fig. S1). Supposedly, the correlative strength of low-frequency spontaneous fluctuations in brain activity is considerably stronger between brain regions within established resting-state networks relative to the strength of correlation between individual networks. A priori we had expected that drugs mainly would modulate connectivity within well-established networks. However, we observed little support for this. Instead, our findings could tentatively be interpreted as if short-term effects of L-dopa and oxazepam are insufficient to cause robust and extensive changes in the underlying neurophysiological basis for resting-state activity, including modulation of slow cortical potentials (local field potentials) believed to be involved in the generation of low-frequency BOLD signal fluctuations (Pizoli et al., 2011).

The absence of changes in local connectivity in the current study contrast earlier findings of L-dopa induced changes of KCC-ReHo (an analogous measure to Cohe-ReHo) were Parkinson patients on medications showed a normalization of KCC-ReHo relative to patients off-medication (Wu et al., 2009). This incongruity could likely be ascribed to differences in cohort and L-dopa dosages, for which the doses, in the study of Wu and associates (2009), were administered on a daily basis, and on average, the doses were four times higher than the single dose given in the current study.

The post hoc comparison that showed an elevated connectivity between vmPFC and bilateral hippocampus in the L-dopa placebo group relative the oxazepam placebo is intriguing. Differences in gender and age between groups were nonsignificant and controlled for, and recruitment procedures as well as time of the year and time of the day when fMRI scanning was performed were similar for both data collection phases. Thus, neither demographical nor scanning differences seem to be plausible explanations for the observed differences in connectivity between the two placebo groups. Rather, a more likely explanation is that the participants in the L-dopa study expected to receive a drug that would potentially increase their sense of excitement, whereas subjects in the oxazepam group expected to be calmed down or even drowsy. A recent study investigating the effects of different resting-state instructions found that attentional stance altered resting-state connectivity in the DMN and frontal areas, supporting the idea that a cognitive context can influence intrinsic activity (Benjamin et al., 2010). Van Kesteren and associates (2010) showed that hippocampal–vmPFC crosstalk supports integration of novel experience and facilitates consolidation of long-term memory, a process that is likely to be more pronounced in mentally aroused subjects. However, speculations based on reversed inference need to be substantiated by, for example, behavioral data on memory performance and subjective reports of arousal. The current study was not aimed or optimally designed for investigating placebo effects per se, so caution should be taken when interpreting these post hoc findings. However, future pharmacological rs-fMRI studies should take potential placebo effects into account.

Several caveats exist for the present study. First, our study employed a cross-sectional design that possibly introduced between the group variance unrelated to effects of drugs, although our experimental groups were well matched with regard to gender and age. Second, the arsenal of methodological approaches to investigate functional connectivity in the resting brain is large, and the selection used in the present study is far from exhaustive. Although three complementary approaches (i.e., seed-based ROI correlation, fALFF, and Cohe-ReHo) were employed, it cannot be ruled out that other methods such as independent component analysis would give additional information on potential drug-related modulations of intrinsic connectivity (Nallasamy and Tsao, 2011). Furthermore, subjects were scanned 1.5 h after drug administration, but there is extensive interindividual variability in the time-to-peak-plasma level, which could attenuate group differences in resting-state activity. The time to peak onset of L-dopa occurs between 0.5 to 4 h after ingestion, with a median value of 1.5 h (half-life 0.75–1.5 h) (Wade et al., 1974). Oxazepam peaks between 0.5 to 8 h after indigestion with an average peak around 2.5 h, (half-life 5–19 h) (Greenblatt et al., 1980).

Future extensions of the present study should correlate behavioral estimates of mood changes with altered resting-state activity, monitoring the plasma-levels, possibly using different drug doses, and study the effects at different time points of the pharmacokinetic cycle.

Conclusions

In sum, it was found that a single administration of oxazepam or L-dopa typically prescribed in clinical settings only has moderate effects on resting-state functional connectivity. It was observed that L-dopa induced increased striatal-frontal connectivity similar to what has been reported previously. However, the overall effect of L-dopa was decreased connectivity, such as within the DMN (PCC/precuneus and vmPFC), between the occipital and temporal areas, and notably between the Am and bilateral frontal gyri. In contrast, oxazepam caused a general increase in connectivity, in particular between the two seed regions in the DMN and parietal, frontal, and cerebellar areas. However, oxazepam-induced decreases in connectivity, for example, between the Am and auditory cortex were also identified. Future studies will be required to clarify the relationship between reported modulations of resting-state activity and psychological changes.

Footnotes

Acknowledgments

Funding/support: This study was supported by the grants from the Swedish Research Council and the Karolinska Institute (P.F.)

Authors Disclosure Statement

No competing financial interests exist.

References

Abadie

, Baron

, Bisserbe

, Boulenger

, Rioux

, Travère

et al. 1992. Central benzodiazepine receptors in human brain: estimation of regional Bmax and KD values with positron emission tomography. Eur J Pharmacol, 213:107–115.

Benjamin

, Lieberman

, Chang

, Ofen

, Whitfield-Gabrieli

, Gabrieli

JDE

et al. 2010. The influence of rest period instructions on the default mode network. Front Hum Neurosci, 4:218.

Biswal

, Yetkin

, Haughton

, Hyde

. 1995. Functional connectivity in the motor cortex of resting human brain using echo-planar MRI. Magn Reson Med, 34:537–541.

Cools

, Frank

, Gibbs

, Miyakawa

, Jagust

, D'Esposito

. 2009. Striatal dopamine predicts outcome-specific reversal learning and its sensitivity to dopaminergic drug administration. J Neurosci, 29:1538–1543.

Costafreda

, Brammer

, David

, Fu

CHY

. 2008. Predictors of amygdala activation during the processing of emotional stimuli: a meta-analysis of 385 PET and fMRI studies. Brain Res Rev, 58:57–70.

Damoiseaux

, Rombouts

SRB

, Barkhof

, Scheltens

, Stam

, Smith

et al. 2006. Consistent resting-state networks across healthy subjects. Proc Natl Acad Sci U S A, 103:13848–13853.

Del-Ben

, Ferreira

, Sanchez

, Alves-Neto

, Guapo

, de Araujo

, Graeff

. 2012. Effects of diazepam on BOLD activation during the processing of aversive faces. J Psychopharmacol, 26:443–451.

Diaconescu

, Menon

, Jensen

, Kapur

, McIntosh

. 2010. Dopamine-induced changes in neural network patterns supporting aversive conditioning. Brain Res, 1313:143–161.

di Martino

, Scheres

, Margulies

, Kelly

, Uddin

, Shehzad

, Biswal

et al. 2008. Functional connectivity of human striatum: a resting state fMRI study. Cereb Cortex, 18:2735–2747.

10.

Fox

, Greicius

. 2010. Clinical applications of resting state functional connectivity. Front Syst Neurosci, 4:19.

11.

Fox

, Raichle

. 2007. Spontaneous fluctuations in brain activity observed with functional magnetic resonance imaging. Nat Rev Neurosci, 8:700–711.

12.

Fox

, Snyder

, Justin

, Vincent

, Corbetta

, Van Essen

, Raichle

. 2005. The human brain is intrinsically organized into dynamic, anticorrelated functional networks. Proc Natl Acad Sci U S A, 102:9673–9678.

13.

Fransson

. 2005. Spontaneous low-frequency BOLD signal fluctuations: an fMRI investigation of the resting-state default mode of brain function hypothesis. Hum Brain Mapp, 26:15–29.

14.

Gospic

, Mohlin

, Fransson

, Petrovic

, Johannesson

, Ingvar

. 2011. Limbic justice--amygdala involvement in immediate rejection in the ultimatum game. PLoS Biol, 9:e1001054.

15.

Greenblatt

, Divoll

, Harmatz

, Shader

. 1980. Oxazepam kinetics: effects of age and sex. J Pharmacol Exp Ther, 215:86–91.

16.

Greicius

, Kiviniemi

, Tervonen

, Vainionpää

, Alahuhta

, Reiss

, Menon

. 2008. Persistent default-mode network connectivity during light sedation. Hum Brain Mapp, 29:839–847.

17.

Greicius

, Krasnow

, Reiss

, Menon

. 2003. Functional connectivity in the resting brain: a network analysis of the default mode hypothesis. Proc Natl Acad Sci U S A, 100:253–258.

18.

Honey

, Bullmore

. 2004. Human pharmacological MRI. Trends Pharmacol Sci, 25:366–374Honey G, Suckling J, Zelay F, Long C, Routledge C, Jackson S, et al. 2003. Dopaminergic drug effects on physiological connectivity in a human cortico-striato-thalamic system. Brain 126:1767–1781.

19.

Kelly

, Zubicaray

, di Martino

, Copland

, Reiss

, Klein

, Castellanos

et al. 2009. L-dopa modulates functional connectivity in striatal cognitive and motor networks: a double blind placebo study. J Neurosci, 29:7364–7378.

20.

Kienast

, Hariri

, Schlagenhauf

, Wrase

, Sterzer

, Buchholz

et al. 2008. Dopamine in amygdala gates limbic processing of aversive stimuli in humans. Nat Neurosci, 11:1381.

21.

Kischka

, Kammer

, Maier

, Weisbrod

, Thimm

, Spitzer

. 1996. Dopaminergic modulation of semantic network activation. Neuropsychologia, 34:1107–1113.

22.

Kiviniemi

, Haanpää

, Kantola

, Jauhiainen

, Vainionpää

, Alahuhta

. 2005. Midazolam sedation increases fluctuation and synchrony of the resting brain BOLD signal. Magn Reson Imaging, 23:531–537.

23.

Kwak

. 2010. Altered resting state cortico-striatal connectivity in mild to moderate stage Parkinson's disease. Front Syst Neurosci, 4:1–25.

24.

Liu

, Yan

, Ren

, Yao

, Kiviniemi

. 2010. Using coherence to measure regional homogeneity of resting-state FMRI signal. Front Syst Neurosci, 4:24.

25.

Martuzzi

, Ramani

, Qiu

, Rajeevan

, Constable

. 2010. Functional connectivity and alterations in baseline brain state in humans. Neuroimage, 49:823–834.

26.

Mccabe

, Mishor

, Filippini

, Cowen

, Taylor

, Harmer

. 2011. SSRI administration reduces resting state functional connectivity in dorso-medial prefrontal cortex. Mol Psychiatry, 16:592–594.

27.

Murphy

, Birn

, Handwerker

, Jones

, Bandettini

. 2009. The impact of global signal regression on resting state correlations: are anti-correlated networks introduced? Neuroimage, 44:893–905.

28.

Nallasamy

, Tsao

. 2011. Functional connectivity in the brain: effects of anesthesia. Neuroscientis, 17:94–106.

29.

Park

, Gu

, Kang

, Shin

, Choi

, Lee

et al. 2010. Integration of cross-modal emotional information in the human brain: an fMRI study. Cortex, 46:161–169.

30.

Paulus

, Feinstein

, Castillo

, Simmons

, Stein

. 2005. Dose-dependent decrease of activation in bilateral amygdala and insula by lorazepam during emotion processing. Arch Gen Psychiatry, 62:282–288.

31.

Peltier

, Kerssens

, Hamann

, Sebel

, Byas-Smith

, Hu

. 2005. Functional connectivity changes with concentration of sevoflurane anesthesia. Neuroreport, 16:285–288.

32.

Pizoli

, Shah

, Snyder

, Shimony

, Limbrick

, Raichle

, Schlaggar

et al. 2011. Resting-state activity in development and maintenance of normal brain function. Proc Natl Acad Sci U S A, 108:11638–11643.

33.

Power

, Barnes

, Snyder

, Schlaggar

, Petersen

. 2012. Spurious but systematic correlations in functional connectivity MRI network arise from subject motion. Neuroimage, 59:2142–2154.

34.

Raichle

, Mintun

. 2006. Brain work and brain imaging. Annu Rev Neurosci, 29:449–476.

35.

Rauch

, Shin

, Wright

. 2003. Neuroimaging studies of amygdala function in anxiety disorders. Proc Natl Acad Sci U S A, 985:389–410.

36.

Rosazza

, Minati

. 2011. Resting-state brain networks: literature review and clinical applications. Neurol Sci, 32:773–785.

37.

Satterthwaite

, Wolf

, Loughead

, Ruparel

, Elliott

, Hakonarson

, Gur

. 2012. Impact of in-scanner head motion on multiple measures of functional connectivity: relevance for studies of neurodevelopment in youth. Neuroimage, 60:623–632.

38.

Tan

, Rudolph

, Lüscher

. 2011. Hooked on benzodiazepines: GABAA receptor subtypes and addiction. Trends Neurosci, 34:188–197.

39.

van den Heuvel

, Hulshoff

PHE

. 2010. Exploring the brain network: a review on resting-state fMRI functional connectivity. Eur Neuropsychopharmacol, 20:519–534.

40.

Van Dijk

, Hedden

, Venkataraman

, Evans

, Lazar

, Buckner

. 2010. Intrinsic functional connectivity as a tool for human connectomics: theory, properties, and optimization. J Neurophysiol, 103:297–321.

41.

van Kesteren

MTR

, Fernández

, Norris

, Hermans

. 2010. Persistent schema-dependent hippocampal-neocortical connectivity during memory encoding and postencoding rest in humans. Proc Natl Acad Sci U S A, 107:7550–7555.

42.

Vincent

, Patel

, Fox

, Snyder

, Baker

, Van Essen

, Zempel

et al. 2007. Intrinsic functional architecture in the anaesthetized monkey brain. Nature, 447:83–86.

43.

Volkow

, Fowler

, Gatley

, Logan

, Wang

, Ding

, Dewey

. 1996. PET evaluation of the dopamine system of the human brain. J Nucl Med, 37:1242–1256.

44.

Wade

, Mearrick

, Birkett

, Morris

. 1974. Variability of L-dopa absorption in man. Aust N Z J Med, 4:138–143.

45.

Weissenbacher

, Kasess

, Gerstl

, Lanzenberger

, Moser

, Windischberger

. 2009. Correlations and anticorrelations in resting-state functional connectivity MRI: a quantitative comparison of preprocessing strategies. Neuroimage, 47:1408–1416.

46.

, Long

, Zang

, Wang

, Hallett

, Li

, Chan

. 2009. Regional homogeneity changes in patients with Parkinson's disease. Human Brain Mapp, 30:1502–1510.

47.

Zou

, Zhu

, Yang

, Zuo

, Long

, Cao

, Wang

et al. 2008. An improved approach to detection of amplitude of low-frequency fluctuation (ALFF) for resting-state fMRI: fractional ALFF. J Neurosci Methods, 172:137–141.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.43 MB

0.27 MB

0.07 MB

0.08 MB